##// END OF EJS Templates
verify: improve progress descriptions...
Steve Borho -
r10906:6ced27d3 default
parent child Browse files
Show More
@@ -1,298 +1,298 b''
1 1 # verify.py - repository integrity checking for Mercurial
2 2 #
3 3 # Copyright 2006, 2007 Matt Mackall <mpm@selenic.com>
4 4 #
5 5 # This software may be used and distributed according to the terms of the
6 6 # GNU General Public License version 2 or any later version.
7 7
8 8 from node import nullid, short
9 9 from i18n import _
10 10 import os
11 11 import revlog, util, error
12 12
13 13 def verify(repo):
14 14 lock = repo.lock()
15 15 try:
16 16 return _verify(repo)
17 17 finally:
18 18 lock.release()
19 19
20 20 def _verify(repo):
21 21 mflinkrevs = {}
22 22 filelinkrevs = {}
23 23 filenodes = {}
24 24 revisions = 0
25 25 badrevs = set()
26 26 errors = [0]
27 27 warnings = [0]
28 28 ui = repo.ui
29 29 cl = repo.changelog
30 30 mf = repo.manifest
31 31 lrugetctx = util.lrucachefunc(repo.changectx)
32 32
33 33 if not repo.cancopy():
34 34 raise util.Abort(_("cannot verify bundle or remote repos"))
35 35
36 36 def err(linkrev, msg, filename=None):
37 37 if linkrev != None:
38 38 badrevs.add(linkrev)
39 39 else:
40 40 linkrev = '?'
41 41 msg = "%s: %s" % (linkrev, msg)
42 42 if filename:
43 43 msg = "%s@%s" % (filename, msg)
44 44 ui.warn(" " + msg + "\n")
45 45 errors[0] += 1
46 46
47 47 def exc(linkrev, msg, inst, filename=None):
48 48 if isinstance(inst, KeyboardInterrupt):
49 49 ui.warn(_("interrupted"))
50 50 raise
51 51 err(linkrev, "%s: %s" % (msg, inst), filename)
52 52
53 53 def warn(msg):
54 54 ui.warn(msg + "\n")
55 55 warnings[0] += 1
56 56
57 57 def checklog(obj, name, linkrev):
58 58 if not len(obj) and (havecl or havemf):
59 59 err(linkrev, _("empty or missing %s") % name)
60 60 return
61 61
62 62 d = obj.checksize()
63 63 if d[0]:
64 64 err(None, _("data length off by %d bytes") % d[0], name)
65 65 if d[1]:
66 66 err(None, _("index contains %d extra bytes") % d[1], name)
67 67
68 68 if obj.version != revlog.REVLOGV0:
69 69 if not revlogv1:
70 70 warn(_("warning: `%s' uses revlog format 1") % name)
71 71 elif revlogv1:
72 72 warn(_("warning: `%s' uses revlog format 0") % name)
73 73
74 74 def checkentry(obj, i, node, seen, linkrevs, f):
75 75 lr = obj.linkrev(obj.rev(node))
76 76 if lr < 0 or (havecl and lr not in linkrevs):
77 77 if lr < 0 or lr >= len(cl):
78 78 msg = _("rev %d points to nonexistent changeset %d")
79 79 else:
80 80 msg = _("rev %d points to unexpected changeset %d")
81 81 err(None, msg % (i, lr), f)
82 82 if linkrevs:
83 83 if f and len(linkrevs) > 1:
84 84 try:
85 85 # attempt to filter down to real linkrevs
86 86 linkrevs = [l for l in linkrevs
87 87 if lrugetctx(l)[f].filenode() == node]
88 88 except:
89 89 pass
90 90 warn(_(" (expected %s)") % " ".join(map(str, linkrevs)))
91 91 lr = None # can't be trusted
92 92
93 93 try:
94 94 p1, p2 = obj.parents(node)
95 95 if p1 not in seen and p1 != nullid:
96 96 err(lr, _("unknown parent 1 %s of %s") %
97 97 (short(p1), short(n)), f)
98 98 if p2 not in seen and p2 != nullid:
99 99 err(lr, _("unknown parent 2 %s of %s") %
100 100 (short(p2), short(p1)), f)
101 101 except Exception, inst:
102 102 exc(lr, _("checking parents of %s") % short(node), inst, f)
103 103
104 104 if node in seen:
105 105 err(lr, _("duplicate revision %d (%d)") % (i, seen[n]), f)
106 106 seen[n] = i
107 107 return lr
108 108
109 109 if os.path.exists(repo.sjoin("journal")):
110 110 ui.warn(_("abandoned transaction found - run hg recover\n"))
111 111
112 112 revlogv1 = cl.version != revlog.REVLOGV0
113 113 if ui.verbose or not revlogv1:
114 114 ui.status(_("repository uses revlog format %d\n") %
115 115 (revlogv1 and 1 or 0))
116 116
117 117 havecl = len(cl) > 0
118 118 havemf = len(mf) > 0
119 119
120 120 ui.status(_("checking changesets\n"))
121 121 seen = {}
122 122 checklog(cl, "changelog", 0)
123 123 total = len(repo)
124 124 for i in repo:
125 ui.progress(_('checking'), i, total=total)
125 ui.progress(_('changesets'), i, total=total)
126 126 n = cl.node(i)
127 127 checkentry(cl, i, n, seen, [i], "changelog")
128 128
129 129 try:
130 130 changes = cl.read(n)
131 131 mflinkrevs.setdefault(changes[0], []).append(i)
132 132 for f in changes[3]:
133 133 filelinkrevs.setdefault(f, []).append(i)
134 134 except Exception, inst:
135 135 exc(i, _("unpacking changeset %s") % short(n), inst)
136 ui.progress(_('checking'), None)
136 ui.progress(_('changesets'), None)
137 137
138 138 ui.status(_("checking manifests\n"))
139 139 seen = {}
140 140 checklog(mf, "manifest", 0)
141 141 total = len(mf)
142 142 for i in mf:
143 ui.progress(_('checking'), i, total=total)
143 ui.progress(_('manifests'), i, total=total)
144 144 n = mf.node(i)
145 145 lr = checkentry(mf, i, n, seen, mflinkrevs.get(n, []), "manifest")
146 146 if n in mflinkrevs:
147 147 del mflinkrevs[n]
148 148 else:
149 149 err(lr, _("%s not in changesets") % short(n), "manifest")
150 150
151 151 try:
152 152 for f, fn in mf.readdelta(n).iteritems():
153 153 if not f:
154 154 err(lr, _("file without name in manifest"))
155 155 elif f != "/dev/null":
156 156 filenodes.setdefault(f, {}).setdefault(fn, lr)
157 157 except Exception, inst:
158 158 exc(lr, _("reading manifest delta %s") % short(n), inst)
159 ui.progress(_('checking'), None)
159 ui.progress(_('manifests'), None)
160 160
161 161 ui.status(_("crosschecking files in changesets and manifests\n"))
162 162
163 163 total = len(mflinkrevs) + len(filelinkrevs) + len(filenodes)
164 164 count = 0
165 165 if havemf:
166 166 for c, m in sorted([(c, m) for m in mflinkrevs
167 167 for c in mflinkrevs[m]]):
168 168 count += 1
169 169 ui.progress(_('crosschecking'), count, total=total)
170 170 err(c, _("changeset refers to unknown manifest %s") % short(m))
171 171 mflinkrevs = None # del is bad here due to scope issues
172 172
173 173 for f in sorted(filelinkrevs):
174 174 count += 1
175 175 ui.progress(_('crosschecking'), count, total=total)
176 176 if f not in filenodes:
177 177 lr = filelinkrevs[f][0]
178 178 err(lr, _("in changeset but not in manifest"), f)
179 179
180 180 if havecl:
181 181 for f in sorted(filenodes):
182 182 count += 1
183 183 ui.progress(_('crosschecking'), count, total=total)
184 184 if f not in filelinkrevs:
185 185 try:
186 186 fl = repo.file(f)
187 187 lr = min([fl.linkrev(fl.rev(n)) for n in filenodes[f]])
188 188 except:
189 189 lr = None
190 190 err(lr, _("in manifest but not in changeset"), f)
191 191
192 192 ui.progress(_('crosschecking'), None)
193 193
194 194 ui.status(_("checking files\n"))
195 195
196 196 storefiles = set()
197 197 for f, f2, size in repo.store.datafiles():
198 198 if not f:
199 199 err(None, _("cannot decode filename '%s'") % f2)
200 200 elif size > 0:
201 201 storefiles.add(f)
202 202
203 203 files = sorted(set(filenodes) | set(filelinkrevs))
204 204 total = len(files)
205 205 for i, f in enumerate(files):
206 206 ui.progress(_('checking'), i, item=f, total=total)
207 207 try:
208 208 linkrevs = filelinkrevs[f]
209 209 except KeyError:
210 210 # in manifest but not in changelog
211 211 linkrevs = []
212 212
213 213 if linkrevs:
214 214 lr = linkrevs[0]
215 215 else:
216 216 lr = None
217 217
218 218 try:
219 219 fl = repo.file(f)
220 220 except error.RevlogError, e:
221 221 err(lr, _("broken revlog! (%s)") % e, f)
222 222 continue
223 223
224 224 for ff in fl.files():
225 225 try:
226 226 storefiles.remove(ff)
227 227 except KeyError:
228 228 err(lr, _("missing revlog!"), ff)
229 229
230 230 checklog(fl, f, lr)
231 231 seen = {}
232 232 for i in fl:
233 233 revisions += 1
234 234 n = fl.node(i)
235 235 lr = checkentry(fl, i, n, seen, linkrevs, f)
236 236 if f in filenodes:
237 237 if havemf and n not in filenodes[f]:
238 238 err(lr, _("%s not in manifests") % (short(n)), f)
239 239 else:
240 240 del filenodes[f][n]
241 241
242 242 # verify contents
243 243 try:
244 244 t = fl.read(n)
245 245 rp = fl.renamed(n)
246 246 if len(t) != fl.size(i):
247 247 if len(fl.revision(n)) != fl.size(i):
248 248 err(lr, _("unpacked size is %s, %s expected") %
249 249 (len(t), fl.size(i)), f)
250 250 except Exception, inst:
251 251 exc(lr, _("unpacking %s") % short(n), inst, f)
252 252
253 253 # check renames
254 254 try:
255 255 if rp:
256 256 if lr is not None and ui.verbose:
257 257 ctx = lrugetctx(lr)
258 258 found = False
259 259 for pctx in ctx.parents():
260 260 if rp[0] in pctx:
261 261 found = True
262 262 break
263 263 if not found:
264 264 warn(_("warning: copy source of '%s' not"
265 265 " in parents of %s") % (f, ctx))
266 266 fl2 = repo.file(rp[0])
267 267 if not len(fl2):
268 268 err(lr, _("empty or missing copy source revlog %s:%s")
269 269 % (rp[0], short(rp[1])), f)
270 270 elif rp[1] == nullid:
271 271 ui.note(_("warning: %s@%s: copy source"
272 272 " revision is nullid %s:%s\n")
273 273 % (f, lr, rp[0], short(rp[1])))
274 274 else:
275 275 fl2.rev(rp[1])
276 276 except Exception, inst:
277 277 exc(lr, _("checking rename of %s") % short(n), inst, f)
278 278
279 279 # cross-check
280 280 if f in filenodes:
281 281 fns = [(lr, n) for n, lr in filenodes[f].iteritems()]
282 282 for lr, node in sorted(fns):
283 283 err(lr, _("%s in manifests not found") % short(node), f)
284 284 ui.progress(_('checking'), None)
285 285
286 286 for f in storefiles:
287 287 warn(_("warning: orphan revlog '%s'") % f)
288 288
289 289 ui.status(_("%d files, %d changesets, %d total revisions\n") %
290 290 (len(files), len(cl), revisions))
291 291 if warnings[0]:
292 292 ui.warn(_("%d warnings encountered!\n") % warnings[0])
293 293 if errors[0]:
294 294 ui.warn(_("%d integrity errors encountered!\n") % errors[0])
295 295 if badrevs:
296 296 ui.warn(_("(first damaged changeset appears to be %d)\n")
297 297 % min(badrevs))
298 298 return 1
General Comments 0
You need to be logged in to leave comments. Login now